4. Methods of Analysis
4.1. Thermal Desorption/Gas Chromatography/Mass Spectrometry (TD/GC/MS)
Several methods of analysis of DEAE and other tertiary amines by different analytical techniques have been published. Gas chromatographic methods use direct injection onto the GC column of underivatized DEAE in aqueous solutions (ASTM D 4983-89; Malaiyanda and Goddard, 1990), underivatized DEAE in organic solutions (Lester and White, 1967), and silylated derivatives of DEAE in organic solution (White and Swafford, 1973); desorption from absorbent gas sampling traps (Fannick, et al, 1983; Visscher, 1990); or direct injection of gas samples into special GC apparatus (Edgerton, et al., 1989). Liquid chromatographic methods rely on formation of colored or radioactive derivatives (Michelot, et al., 1983), as do spectrophotometric and colorimetric methods (Larrick, 1963; Miller, et al., 1967). Only the publications of Fannick, et al (1983) and Visscher (1990) DEAE with the analysis of DEAE in museums, and both use gas adsorbent trap – GC methods. The GC methods are most sensitive and convenient. For this project a thermal desorption (TD) method of sample introduction onto a GC column was chosen. This TD method is a modification of one that has been used for several years at CCI for GC/MS analysis of volatile compounds emitted by paints and adhesives.
The TD method involves two steps which are carried out in a Thermal Desorption Unit (TDU). In the first step, the sample preparation or loading mode, a sample of either gas or liquid is injected through a septum into a sample tube, an empty 1/4” OD glass tube, contained in a temperature controlled tube chamber, then the tube chamber sample preparation heating program is activated. Carrier gas flowing through the sample tube, sweeps the volatilized sample out of the sample tube onto an adsorbent tube which is a 1/4” OD, 1 mm ID glass tube packed with adsorbents (Carbotrap 301) that adsorb efficiently organic molecules having a size greater than equivalent to alkanes with about 2 to 3 carbon atoms (i.e., compounds with molecular weights greater than about 30-45 atomic mass units, amu). Thus, water, oxygen, carbon dioxide, and methanol, for example, pass through the adsorbent tube without being adsorbed, whereas larger molecules such as DEAE (MW =117 amu) are completely adsorbed in the adsorbent tube. By this means, small amounts of DEAE can be concentrated in the adsorbent tube by injecting large volumes of methanol solution (e.g., 5 μL) into the sample tube.
In the second step, the desorption mode, the adsorbent tube with the adsorbed analyte is placed in the tube chamber. Carrier gas flow through the tube is switched from the path to the adsorbent tube to the path through the GC column. The temperature of the tube chamber is raised rapidly to 330 C to desorb the sample from the adsorbent tube and the desorbed sample is swept through the GC where separation and subsequent detection with the mass spectrometer occurs.
One advantage of this TD sample injection method is that concentration of the DEAE in methanol extracts by evaporation of the methanol is not necessary, so there is no loss of DEAE by simultaneous evaporation, or azeotropic distillation. Another advantage is that, since most of the methanol passes through the adsorbent tube, no methanol is injected onto the analytical column when the sample is desorbed so there is no large background signal in the chromatogram due to methanol solvent. Contaminants in the methanol are concentrated so high purity solvent is required.
All swabs were extracted with methanol added directly to the sample vial, typically 0.5 mL which was just sufficient to cover the swabs in the vials For gas chromatographic analysis, samples of the extract were removed from the vial by a syringe pierced through the septum in the vial cap then injected into the thermal desorption unit.
Powder and flake samples stored in vials were extracted, dissolved, or dispersed in methanol, typically 0.25 to 0.5 mL, by addition of methanol directly to the vial. Samples of these liquids were removed from the vial by a syringe pierced through the septum in the vial cap.
To prevent loss of volatile DEAE or its reaction products, all samples collected during Williams’ visit were kept in septum sealed glass vials, except for the brief period when methanol was added to the vial Also no methanol solutions or mixtures were subjected to evaporation to concentrate the samples.
The TD/GC/MS apparatus and instrument settings were as follows:
Gas Chromatograph/Mass Spectrometer Hewlett Packard (HP) 5870 GC with a HP 5970B Mass Selective Detector (manual dated January 1986) controlled by a HP 59970C MS ChemStation with HIP 59974J GC/MS Software (revision 3.1.1, copyright 1986) using a HP 9133H Disc Drive.
GC Column: DB-WAX, 30 m x 0.25 mm ID x 0.25 μm film thickness (J&W Scientific, P/N 122-7032) received on 16/8/94.
GC Oven: 45°C for 2 min, then to 200°C at 20oC/min and hold.
Thermal Desorption Unit: Dynathenn Analytical Instruments, inc. Thermal Desorption Unit (TDU) Model 890/891 from Supelco, Inc.
TDU Adsorbent Trap: Carbotrap 301 Multibed Thermal Desorption Tube, 1 mm ID (Supelco Catalog No. 2-0354).
Split Ratio at TDU exit. 60:40 (column:vent)
TDU Temperature Conditions:
Preparation (loading) mode: initial approx 45°C, final 300°C, 4 min hold.
Desorption mode: initial: approx 45°C, final: 330°’C, 4 min hold.
Injection Volume (typical): 5 μL of methanol extract or solution using a Hamilton 701 syringe.
4.2. Fourier Transform Infrared (FTIR) Spectroscopy
A few individual particles from powdery scrapings or excised chips and flakes were analyzed by Fourier transform infrared spectroscopy using a Spectra-Tech IR-Plan lit microscope
interfaced to a Bomem MB-120 FTIR Spectrometer.
Samples were prepared in a low pressure diamond anvil sample cell from High Pressure Diamond Optics by placing a particle on one anvil, assembling the cell, then squeezing the sample by applying pressure until the sample was about 10 μm thick. The diamond cell was opened and the anvil with the sample stuck to it was mounted in the IR microscope for spectroscopy of circular areas of the squeezed samples measuring 100 μm in diameter, using clear areas of the diamond anvil immediately adjacent to the sample for background spectra for each sample spectrum.
The typical sample size was about 10 μm thick by 100 μm in diameter. Assuming that the density of the sample is about 1 gm/cm2, the weight of the sample analyzed can be calculated:
10 μm x π x (50 μm)2=l0x 3.14 x 50 x 50 = 78450 μm3
=7.8E4 μm3 x (1 cm/ 104 μm)3
=7.8E-8 cm3 x 1 gm(sample)/cm3
FTIR spectroscopy is usually capable of detecting components in mixtures that comprise 1% or more of the total sample weight when there is no overlap of absorption bands for the components. DEAE has absorption bands that are not masked by absorption bands of resin and acrylic varnishes or oil and protein paint media. Thus the limit of detection for DEAE in painting samples by this FTIR microspectroscopic technique should be about 1% of 78 ng or 0.78 ng DEAE.
When the sample is on the diamond anvils it can be treated with reagents by placing drops of reagents on the sample as it rests on the diamond, allowing these to react and then evaporate, then acquiring spectra of the dry reaction products.
This procedure has been used for years at CCI to remove lead and calcium carbonates from samples by adding hydrochloric acid (which reacts with the carbonate to produce carbon dioxide gas and IR transparent calcium or lead chlorides) or to remove silica and silicates by adding hydrofluoric acid (which produces volatile silicon tetrafluoride and IR transparent metal fluorides).
The judicious addition of hydrochloric acid or sodium or ammonium hydroxide in various sequences can also be used to probe carboxylic ester, acid, and salt functional groups. For example, the presence of a carboxylate salt (e.g., zinc stearate) can be confirmed by observing the shift of the Zn carboxylate absorption from 1540 cm-1 to 1710 cm-1 for carboxylic acid when hydrochloric acid is added. Subsequent addition of sodium hydroxide causes the 1710 cm-1 peak of acid to disappear while the 1580-1540 cm-1 peak for sodium carboxylate salt (soap), appears.
The effects of DEAE on the IR spectra of samples was investigated by placing a drop of DEAE on the sample after an initial spectrum of the untreated material had been obtained. These reactions were observed using a stereomicroscope. The process of evaporation of the DEAE was observed and in some cases hastened by the heat from the illuminating lamps Reaction products with the DEAE were subsequently treated with water, mineral acids, and alkalis to observe their reactions. Spectra of the reaction products were acquired.
4.3. Freeze fracture experiments and scanning electron microscopy (SEM)
5 mm wide strips were cut from the test paintings using a sharp scalpel yielding test pieces about 5 mm wide x 50 mm long. The test pieces were immersed in liquid nitrogen (LN2). Immediately after removal from the LN2 the test pieces were bent around a 5 mm diameter metal rod to fracture the paint layer. The support layer did not fracture and was cut with a scalpel. Two pieces of each test piece resulted. One piece has been stored in a glass vial as a control. The other piece has been suspended by a thread in a vial over 1 mL of pure DEAE. These were to be examined by scanning electron microscopy (SEM).